Research On The Design And Application Of Credit Scoring System For Credit Card Customers Based On AHP-Logistic Model

At present,China's credit card business is developing rapidly.The scale of card issuance is increasing.The credit card business has gradually become an important source of profit for commercial banks.The credit card business still has a large market waiting for development in China.The future development potential is huge.With the rapid development of credit card business,the credit card risk problem has become increasingly serious,including credit card fraud,malicious overdrafts and theft of credit cards and other criminal events.This directly affects the further development of the credit card business.Therefore,how to improve credit card risk management,including the establishment of a scoring model and a scientific risk management system has become the first problem to be solved in the development of credit card business in China.At present,China's financial market has entered a free market state.Major banks have focused their competition on retail banking business.Among them,the credit card business is particularly fiercely competitive.The characteristics of credit card business are high risk and high returns,so how to control risk? To the minimum,to maximize profits is a key issue in China's credit card business.This article is based on commercial bank credit card users apply for the scorecard data,based on combining AHP customer stratification results to establish an improved AHP-Logistic model,systematic evaluation of regular scorecard method and qualitative data statistical analysis to enhance business The scientific level of daily bank credit card customer risk management.The innovation of this article is:First,using the existing customer segmentation model and combining credit risk management objectives,scientifically select customer behavioral risk indicators,and apply the AHP Analytic Hierarchy Process(AHP)to weight the 19 indicators in the three categories of personal basic conditions,economic conditions,and credit conditions.Sorting,using AHP to build a hierarchical evaluation model for customer risk measurement.Secondly,on the basis of stepwise regression,the influence factors of the credit card risk level were screened out,and the AHP customer stratification result was introduced into the Logistic model as a new explanatory variable,and an improved AHP-Logistic credit card scoring model was proposed.Combining with empirical analysis,the related theoretical problems such as parameter estimation and prediction test were discussed.The accuracy and stability of the improved AHP-Logistic scoring model were better thanthose of the single AHP model and the traditional Logistic model.Finally,from the above improved AHP-Logistic model,we have obtained a more accurate customer credit evaluation model.At the same time,we can convert the obtained model parameter values and arrive at a corresponding score,making the model practical.Applications.At the end of the article,the conclusions drawn from the study of this article are summarized,and where further research is needed should be improved.